为了降低内容寻址网络CAN(content-addressable network)节点间的延迟,建立了数学模型,引入求静态图最短路径的Dijkstra算法,并以重叠网络中的节点为图的顶点,相邻节点以边连接,相邻节点间的延迟为边的权值构建节点间的动态延迟图.本文的算法能在任意两个节点间的多条路径中找到一条延迟最小的路径.使用P2Psim对该方法和选择延迟最小的邻居节点作为下一跳的方法进行对比测试,结果表明本文的方法能更有效地降低CAN中节点间的延迟.
Content-addressable network(CAN) is a method of locating resources in structured peer-to- peer network. To reduce the delay between CAN nodes, a mathmetic model and a dynamic graph is built, Dijkstra algorithm is applied on the dynamic graph to find the smallest delay between CAN nodes. Our method is compared with previous method by simulator. Experiment results show that our method could reduce the delay between CAN nodes efficiently.